Familiarity Does Indeed Promote Attraction in Live Interaction … · 2011. 4. 3. · Familiarity...

Familiarity Does Indeed Promote Attraction in Live Interaction

Harry T. Reis, Michael R. Maniaci, Peter A. Caprariello,

The University of Rochester

Paul W. Eastwick

Texas A & M University

and

Eli J. Finkel

Northwestern University

Direct correspondence to: Harry T. Reis, Ph.D. Department of Clinical and Social Sciences in Psychology University of Rochester Box 270266 Rochester, NY 14627 Voice: (585) 275-8697 Fax: (585) 273-1100 E-mail: [email protected]

1

Abstract

Does familiarity promote attraction? Prior research has generally suggested that it

does, but a recent set of studies by Norton, Frost, and Ariely (2007) challenged that

assumption. Instead, they found that more information about another person, when that

information was randomly selected from lists of trait adjectives, using a trait evaluation

paradigm, promoted perceptions of dissimilarity and hence disliking. The present

research began with the assumption that natural social interaction involves contexts and

processes not present in Norton et al.'s research, nor in the typical familiarity experiment.

We theorized that these processes imply a favorable impact of familiarity on attraction.

Two experiments are reported using a live interaction paradigm in which two previously

unacquainted same-sex persons interacted with each other for varying amounts of time.

Findings strongly supported the "familiarity leads to attraction" hypothesis: The more

participants interacted, the more attracted they were to each other. Mediation analyses

identified three processes that contribute to this effect: perceived responsiveness,

increased comfort and satisfaction during interaction, and perceived knowledge.

2

Among the core concepts of interpersonal attraction is the principle of familiarity.

According to Berscheid and Regan, for example, "the familiarity principle of attraction is

perhaps the most basic of the [general principles of attraction]" (2005, p. 177). Similarly,

Ebbsen, Kjos, and Konecni (1976, p. 505) conclude that "most positive interpersonal

relationships result from frequent face-to-face contacts." These conclusions follow from

the many studies, both correlational and experimental, that have supported a link between

familiarity – defined as the degree of exposure that one person has to another person –

and attraction to other persons. Consistent with this definition, the familiarity effect on

attraction is typically explained in terms of the mere exposure effect (Zajonc, 1968, 2001)

– that repeated exposure to a stimulus increases liking for that stimulus – although, as we

discuss later, we believe that familiarity effects in social interaction involve more

interpersonal processes.

A recent article by Norton, Frost, and Ariely (2007) challenged this conclusion.

These authors proposed that familiarity tends to breed dislike, because familiarity, which

they defined as acquiring more information about another person and typically

operationalized in terms of acquiring random bits of information about that person, is

likely to disconfirm assumptions about another person's similarity to oneself. Ambiguity,

on the other hand, which they defined as the absence of information, was said to breed

liking, because it facilitates the assumption that the other is similar to oneself, which

makes it easier to imagine liking the other. Their paper reported a series of clever

experiments (described below) that supported their reasoning.

The present manuscript is based on our belief that although Norton et al.'s (2007)

findings may be internally valid, they misrepresent the typical operation of familiarity in

acquaintanceship based on live interaction. That is, their model and research defines and

3

operationalizes familiarity in terms of the amount of information that one has about

another person. Although the acquisition and assessment of information is surely part of

developing acquaintanceships, we propose that increasing familiarity in interactive

relationships is a considerably more complex process, involving responsive interaction

and affective experience, as well as other forms of interpersonal influence (Kelley et al.,

1983). Moreover, their study presented information to participants that had been selected

randomly from lists of positive and negative traits. As discussed below, we suspect that

this is not representative of information exchange in real-world social interaction. Finally,

we propose that knowledge gained about another person is assimilated and interpreted

differently when it is acquired in the ebb-and-flow of interaction than when it is acquired

acontextually.

Norton et al.'s (2007) research used paradigms that are relatively similar to

previous experiments on familiarity, in the sense that information was presented to

participants in a very decontextualized manner. To our knowledge, no experiments have

examined the "familiarity-leads-to-attraction" effect in contexts involving actual

interaction; most experiments have followed the example of mere exposure studies,

presenting stimuli such as names, faces, or trait information at varying frequencies.

(Familiarity effects have been examined in natural settings [e.g., Berg, 1984; Shook &

Fazio, 2008], as described below, but because none of these studies was a true

experiment, their interpretations are potentially ambiguous.) There is an important

distinction, we believe, between trait evaluation paradigms, in which participants

evaluate static information about a person they will never meet, and live interaction

paradigms, in which people interact in real time, acquire information contextually, and

both evaluate and are evaluated by the partner. In other words, natural interaction differs

4

in several important respects from thinking critically about lists of information. In fact,

existing evidence suggests that information may be processed differently – that is, more

holistically – when it is embedded in the ebb and flow of natural interaction. For

example, in two laboratory experiments reported by Eastwick, Finkel, and Eagly (2010),

participants, on the basis of written profiles, preferred ideal to nonideal romantic partners.

After a live interaction, however, this preference disappeared, because interaction

facilitated more holistic, contextual interpretation of trait information, as traditional

models of person perception have long assumed (Asch, 1946). Thus, we believe that the

present experiments represent a more ecologically valid test of the familiarity-attraction

hypothesis than prior experiments. Building on traditional familiarity-attraction research

and notwithstanding Norton et al.'s (2007) novel contribution to that literature, we

propose that in the context of actual social interactions, familiarity is associated with

increasing attraction. This paper reports two experiments supporting this position.

Explaining Why Familiarity Breeds Liking

Familiarity effects are often couched in terms of the mere exposure effect, so we

begin with a brief review of that literature. Researchers have studied diverse phenomena

relevant to the mere exposure effect since it was first postulated (Zajonc, 1968).

Bornstein's (1989) meta-analysis documents significant mere exposure effects with

regard to several types of stimuli: sounds, ideographs, nonsense words and symbols,

drawings, photographs, words and names, polygons, objects, and persons. Several of

these studies were concerned with interpersonal relations. The classic study in this area is

Moreland and Beach (1992). In that research, four female confederates entered a

classroom in a manner visible to other students either zero, five, ten, or fifteen times over

the course of a semester; they did not interact with the students in the class. Afterwards,

5

students were asked to rate the confederate on various dimensions. The more often she

had been seen by students, the more she was liked and rated positively on various

dimensions. Other studies have shown that the more frequently another person has been

seen, the more participants rated that person positively and wanted to interact with him or

her (Brockner & Swap, 1976); the longer participants were exposed to another persons'

ideas, the more they liked that person (Brickman, Meyer, & Fredd, 1975); the more

familiar a negotiation partner, the more willing people are to reach compromise solutions

(Druckman & Broome, 1991); and the more familiar a public figure's picture or name, the

more likeable that person is perceived to be (Harrison, 1969). In a particularly clever

study, Mita, Dermer, and Knight (1977) showed that people preferred their own facial

image when reversed over true facial images (because the reversed image is more

familiar due to everyday grooming). Familiarity effects on liking have been observed in

both Western and Asian cultures (Heine & Renshaw, 2002) and at least one researcher

has suggested that familiarity provides a necessary context for imitation in babies, an

important component of cognitive development (Parker-Rees, 2007).

Familiarity effects are important in their own right, and also because they have

been used to explain why proximity (also called propinquity), another venerable factor in

the attraction literature, predicts attraction. Many studies have shown that proximity

predicts liking (e.g., Back, Schmukle, & Egloff, 2008; Festinger, Schachter, & Back,

1950; Nahemow & Lawton, 1975; Segal, 1974). People tend to encounter others in close

proximity more often, hence they become more familiar.

Four general explanations have been offered for why familiarity predicts

attraction. The first involves classical conditioning. Because most social experience is

mildly positive in affective tone, or at least not aversive (Clark & Watson, 1988; Denrell,

6

2005; Reis & Gable, 2003), more frequently encountered others become paired with

positive affect. Second, for evolutionary reasons, novel stimuli foster uncertainty and

wary reactions (Lee, 2001; Orive & Gerard, 1987), which tends to diminish once repeated

exposure has shown the novel stimulus to be benign. The wary response to strangers is

common in human and other animals (Rajecki, 1985). Third, familiar stimuli tend to be

processed perceptually and cognitively in a more fluent manner, and fluency tends to be

experienced in relatively positive affective ways (Bornstein & D'Agostino, 1994;

Winkielman, Schwarz, Fazendiero, & Reber, 2003).

The fourth explanation, which extends the classical conditioning argument, is

more directly germane to the present research. Familiarity, or repeated exposure, creates

opportunities for interaction. All other things being equal, as mentioned above, positively

toned interactions are more common than aversive interactions, suggesting that such

opportunities are more likely than not to lead to rewarding social experiences and

favorable impressions (Denrell, 2005). Because the "rewards others provide" are a central

motivating factor underlying social interaction (Berscheid & Walster, 1969) and liking

(Berg, 1984), greater familiarity implies that rewarding interactions have taken place in

the past. In other words, more often than not, interaction tends to be experienced on the

positive side of hedonic neutrality, creating the desire for further interaction. Many social

processes contribute to this tendency – for example, politeness norms, the tendency to

identify common interests, and basic affiliative drives (e.g., Baumeister & Leary, 1995).

There are some circumstances in which familiarity does not breed liking, but

instead breeds contempt. For example, Ebbeson et al. (1976) showed that greater

familiarity with obnoxious others led to greater dislike for them, consistent with findings

that repeated exposure to unpleasant stimuli does not enhance liking (e.g., Perlman &

7

Oskamp, 1971). If familiarity creates opportunities for interaction, then to the extent that

those interactions are aversive, familiarity should have negative impact. This is consistent

with response facilitation models of social experience: that repeated contact increases the

likelihood of the dominant (i.e., most predisposed) response (Zajonc, 1965). This

explanation is also consistent with research on the contact hypothesis. Although it was

first assumed that intergroup contact of any sort would reduce prejudice and outgroup

stereotyping, years of research have shown such contacts can make matters worse if the

conditions of such contact exacerbate preexisting dislike and suspicion (Amir, 1976). On

the other hand, intergroup contact under circumstances that foster cooperative interaction

tends to lessen prejudice, consistent with our reasoning. In fact, conditions amenable to

fostering intimate friendship appear to be especially effective in lessening prejudice and

discrimination (Pettigrew, 1998).

Might Familiarity Impair Liking? Norton et al.'s (2007) Research

As mentioned above, Norton et al. (2007) argued that "less is more" – that is, that

as more information about another person is acquired, the likelihood of finding evidence

of dissimilarity increases, which will engender decreased liking. As they concluded their

article, "knowing more means liking less" (2007, p. 103). This account follows from their

finding that people anticipate initial interactions with relatively favorable expectations,

because ambiguity allows for self-serving inferences about the other's opinions and traits.

Three experiments (one conducted in an online dating website, the other two conducted

around the MIT campus, omitting mention of dating or the target's sex) reported in their

article relied on a similar method. Participants were presented varying amounts of trait

information about a potential interaction partner; for example, in one study participants

received a list of 4, 6, 8, or 10 positive and negative traits randomly selected from Asch's

8

(1946) seminal study of central and peripheral traits in impression formation. As

predicted by Norton et al.'s theorizing, the more traits participants received, the less they

anticipated liking the target person. In one additional study, one group of participants in

an online dating service described their expectations about a forthcoming date, as well as

how much they knew in advance about the date. A second group was asked the same

questions after a first date (with appropriate changes in tense). As expected, knowledge

ratings were higher in the post-date condition, but ratings of liking and perceived

similarity were lower1.

We accept these findings, and their accompanying theorizing, as far as they go.

But for the reasons spelled out next, we believe that they may misrepresent the impact of

familiarity effects in actual social interactions. Several lines of reasoning lead us to this

belief. First, the only one of their studies that involved interaction with another person

involved online dating, which may be a special case that cannot be generalized to other

forms of attraction and interaction. Online dating emphasizes evaluation, because

participants typically choose among many alternative partners on the basis of

information-laden profiles provided on the website (Sprecher, Schwartz, Harvey, &

Hatfield, 2008). Furthermore, dating emphasizes the sexual component of attraction in a

way that ordinary social interaction does not. Berscheid and Regan (2005) explain that

familiarity tends to diminish sexual attraction (opposite to its effect on nonsexual

attraction), perhaps because high-arousal emotions such as sexual arousal depend to some

extent on novelty (e.g., Berscheid & Ammazzalorso, 2001; Reis & Aron, 2009). For

example, Wolf (1995) found that sexual attraction tends to be lower in humans raised

together, an instance of the Westermarck hypothesis, which is thought to reflect

evolutionary pressures related to incest taboos. Also, ever since the first computer dating

9

study (Walster, Aronson, Abrahams, & Rotman, 1966), researchers have recognized that

impression formation in first dates is dominated by physical attractiveness in a way that

may not apply to other kinds of acquaintanceships.

Second, and more importantly, lists of abstract traits, as used in trait evaluation

paradigms, present information in a very structured yet decontextualized manner which

seems unlike the way in which people learn about others in everyday social interaction.

For one thing, people in real conversations do not reveal information randomly. Rather,

they disclose information that fits with the ebb-and-flow of the conversations, as well as

with their goals (e.g., self-presentation) for that conversation (Miller & Read, 1987). For

a second reason, trait information usually comes wrapped around anecdotes and

explanations, which make varying interpretations likely. For example, it seems

meaningfully different to learn that a partner is "introverted" than to learn that "I felt

hesitant when I went to my first campus party." Contextualization of information about

others – accomplished, for example, by including individuating information about a

person along with traits – can alter person perception, as has been shown in studies of

social interaction with new acquaintances (Eastwick et al., 2010) or stereotyping (Smith,

Miller, Maitner, Crump, Garcia-Marques, & Mackie, 2006). Furthermore, there are

qualitative differences between live interaction and meeting someone "on paper."

Notably, live interaction affords more holistic processing of information about the other

(Eastwick et al., 2010), whereas trait lists would favor more analytic processing of trait

information. This difference helps explain why preferences expressed in response to real

interaction with potential partners in a speed-dating context may diverge from self-

reports of ideal partner preferences (Eastwick & Finkel, 2008).

10

Third, in conversation, people usually seek out commonalities – other persons that

both parties know, common interests, and similar experiences both have had (Clark,

1996; Stasser, Taylor, & Hanna, 1989). This allows conversants to pursue common

ground, a key component of socially shared cognition (e.g., Clark & Brennan, 1991;

Hardin & Higgins, 1996). Thus, familiarity often leads to the perception of increased

similarity (e.g., Moreland & Zajonc, 1982).

Furthermore, familiarity in everyday social experience involves more than the

acquisition of information. Another source of knowledge involves responsiveness – how

partners respond to each other. Positive responses – for example, support,

encouragement, humor, and engagement – tend to generate affinity whereas uninterested

or critical responses tend to foster indifference or animosity (e.g., Curtis & Miller, 1986).

Moreover, extensive research shows that responsiveness encourages mutual self-

disclosure (see Reis & Patrick, 1996, for a review), creating interactions that let partners

know and feel known by each other (Reis, Clark, & Holmes, 2004). This point is

consistent with research on processes such as self-verification (Swann, 1990), intimacy

(Reis & Shaver, 1988), and responsiveness goals (Canevello & Crocker, in press). These

and many other studies suggest that perceived understanding by others contributes to

feeling accepted, valued, and liked, factors that contribute to reciprocity of liking (Kenny,

1994; Newcomb, 1961). Thus, social interaction (at least when it goes reasonably well)

allows for the interchange of self-disclosure and responsiveness that fosters both

knowledge of the other and the feeling of being known by the other.

Finally, and more generally, familiarity fosters feeling comfortable and safe with

others, which also contributes to liking (Baumeister & Leary, 1995; Floyd, 2006;

11

Winkielman et al., 2003). Attachment figures, for example, are almost without exception

familiar, well-liked others (Mikulincer & Shaver, 2007).

Combining these various considerations points to an important conceptual

distinction. Kruglanski et al. (2000) have distinguished assessment mindsets, which

emphasize analytical reasoning in which the target is evaluated relative to alternatives,

from locomotion mindsets, which emphasize the commitment of self-regulatory resources

to the initiation and maintenance of desired actions. Under locomotion, individuals are

more likely to respond to the other spontaneously, in a manner that facilitates smooth,

rewarding social interaction. Consistent with this reasoning, Kumashiro, Rusbult,

Finkenauer, and Stocker (2007) demonstrated that locomotion orientations were

associated with greater support, affirmation, and relationship well-being, whereas

assessment orientations undermined these qualities. Extensive research based on a closely

related conceptual distinction, differentiating deliberative and implemental mindsets (see

Bargh, Gollwitzer, & Oettingen, 2010; Gollwitzer & Bayer, 2006, for reviews), similarly

suggests that the former may stress purely diagnostic reasoning about the other, whereas

the latter may energize efforts to achieve more harmonious, enjoyable interactions.

We reasoned that Norton et al.'s (2007) trait evaluation paradigm seems likely to

have induced an assessment mindset, in which the individual asks, "Am I interested in

this person?" Because they received only a list of trait adjectives, participants could only

answer the researchers' question by considering the relative merits of those traits. Vis-à-

vis interaction between the evaluator and the target of the evaluation, locomotion is

irrelevant, because no relationship goals could be pursued. On the other hand, actual

social interactions are more likely to induce a locomotion mindset, in which the

individual aims to interact in as comfortable and rewarding a manner as possible.

12

Engaging in social interaction requires committing attention and other psychological

resources to the pursuit of a smooth, non-awkward, and pleasant social experience. At the

same time, this orientation would downplay critical assessments, which would interfere

with spontaneity and social engagement.

This reasoning is supported by the results of an experiment by Snyder and

Haugen (1995), who compared the knowledge functions (acquiring stable impressions of

the other) and adjustive functions (engaging in a smooth, pleasant interaction) of getting-

acquainted conversations. Importantly, they found that when no instructions are provided

in such conversations, adjustive functions predominated over knowledge functions. Our

reasoning is also consistent with a recent report by Frost, Chance, Norton and Ariely

(2008), who demonstrated that an online dating service that allowed potential daters "to

acquire experiential information by exploring a virtual environment in interactions

analogous to real first dates (such as going to a museum)" (p. 51) produced greater liking

than online dating sites that emphasized what they called searchable attributes

(commodities and attributes that can be described succinctly). Thus, both sets of findings

support the conceptual implications of the distinction between trait evaluation and live

interaction paradigms that we are proposing.

For these reasons, we expected that in the context of actual social interactions,

familiarity would lead to increased liking, as earlier studies have suggested. Additionally,

we expected that familiarity would increase participants' knowledge about each other (as

shown by Norton et al., 2007), as well as the perception of being known by the other, and

that these knowledge increases would be associated with greater liking, as the theories

reviewed above suggest, rather than with decreased liking, as Norton et al., 2007, found.

The Present Research

13

This manuscript reports two experiments using the live interaction paradigm

testing the hypothesis that greater familiarity would predict greater attraction. In the first

experiment, in face-to-face conversations, pairs of previously unacquainted participants

discussed either 2 or 6 items taken from validated experimental tasks for generating

closeness between strangers. These items have been demonstrated to allow conversations

to unfold in a way that encourages mutual self-disclosure and supportive responses. In the

second experiment, same-sex strangers were paired and randomly assigned to engage in

online chats with each other 1, 2, 4, 6, or 8 times. Taken as a set, these two experiments

control for factors that may be associated with familiarity in real-world interactions –

Study 1 controls for the topic of conversation whereas Study 2 controls for various cues

that are visible when people interact face-to-face. In addition to examining the effect of

familiarity on attraction, we also examined the mediating role of two knowledge-related

variables (how much knowledge participants felt that they had about their partners and

how much knowledge participants felt that their partners had about them), perceived

responsiveness, and comfort/satisfaction.

Both experiments were conducted with same-sex pairs because, as mentioned

above, interactions between previously unacquainted opposite-sex strangers are often

dominated by concerns about dating and sexual attraction2. We were interested in

studying the effects of familiarity on more general processes of attraction and friendship

formation.

Study 1 Method

Participants

We recruited 56 participants for the current study. Because we were primarily

interested in studying initial interactions, only participants who reported (on a

14

questionnaire at the end of the study) that they had not previously met their interaction

partner were retained for analyses. Two participants who reported not knowing their

partner were retained for analyses even though their interaction partners did not indicate

likewise; thus the final N was 48 participants (25 women) comprising 25 different same-

sex dyads. Four participants completed the study for course credit, whereas the remaining

participants were recruited using flyers posted around campus and were paid $6.

Participants' mean age was 19.31 (SD = 1.09), ranging from 18 to 22.

Measures

Attraction toward the partner was the average of four items (alpha = .79):

perceived similarity (assessed on a 7-point scale, anchored by Not at All and Extremely),

liking for the partner and desire to have the partner as a friend (assessed on 7-point

scales, anchored by Neutral and An exceptional amount), and Inclusion of Other in the

Self (IOS; Aron, Aron, & Smollan, 1992). The IOS was translated to a 7-point scale, with

higher numbers denoting greater inclusion in the self.

Perceived responsiveness was based on a scale developed by Reis (2006), but

modified to reflect new acquaintance. The twelve items are listed in Appendix A. Scores

represented the average of all items (alpha = .93), all of which were assessed with 7-point

scales ranging from Not at all true to Completely true.

Procedure

Two same-sex participants arrived at the laboratory for each experimental session.

After obtaining consent, experimenters brought the two participants to an interaction

room, motioned for them to sit across a table from each other, and explained that they

would be having a short conversation. The experimenter then held up a set of either 2 or 6

cards, each containing a question that the participants were to ask and answer for each

15

other. One participant (“partner A”) would ask the question, then the other participant

(“partner B”) would answer for 30 seconds (using a clock on the wall to keep time).

Then, partner B would ask the same question of partner A, and partner A would respond

for 30 seconds. This process continued until all questions had been asked and answered

by both participants.

Participants were randomly assigned to the 2- or 6-card condition. The questions

on five of the cards were taken from the Relationship Closeness Induction Task (RCIT;

Sedikides, Campbell, Reeder, & Elliot, 1999): “What are your hobbies?,” “What would

you like to do after graduating from Northwestern?,” “What is something you have

always wanted to do but probably never will be able to do?,” “If you could change

anything that happened to you in high school, what would that be?,” and “What is one

thing about yourself that most people would consider surprising?” The question on the

other card came from Aron, Melinat, Aron, Vallone, and Bator's (1997) experimental task

for generating closeness between strangers: “If a crystal ball could tell you the truth about

yourself, your life, your future, or anything else, what would you want to know?”

Participants in the 6-card condition responded to all six questions (which the

experimenter shuffled before handing to partner A), whereas participants in the 2-card

condition responded to two questions randomly selected from the set of six.

The experimenter was not present in the room during the interaction. Immediately

after the final questions had been answered by both participants, the experimenter

reentered the room, took the participants to separate rooms, and administered a

questionnaire containing the attraction and perceived responsiveness measures. Finally,

participants were debriefed and dismissed.

Study 1 Results

16

Because individual responses were nested within dyads, producing

interdependence, the primary unit of analysis was the dyad (Kenny, Kashy & Cook,

2006). Partners' responses were correlated, r(23) = .22, ns, for the attraction measure, and

r(23) = .49, p<.02, for the perceived responsiveness measure. We therefore used the

average of both partners' responses for the primary analyses, except in the two dyads

where measures were retained for only one of the two partners (see Method). In these two

dyads, we used the ratings from the participant who reported not knowing the other

participant in advance3.

Preliminary analyses found no main effects or interactions involving the dyad's

sex, Fs(1, 21) < 1.95, ns, so we did not include sex in the primary hypothesis tests.

Consistent with experimental instructions, in the 2-card condition, participants conversed

for 129 seconds, on average; in the 6-card condition, their conversations averaged 394

seconds; t(13) = 10.93, p<.001.

Analysis of the attraction composite yielded a significant mean difference, t(23) =

2.23, p<.04; d = .91. In the 2-card condition, the mean level of attraction was 3.55 (SD =

.75); in the 6-card condition, the mean level of attraction was 4.17 (SD = .61). Thus,

additional conversation led to increased attraction, as we predicted.

Additional interaction led to marginally higher ratings of perceived

responsiveness, as expected, t(23)=1.83, p<.09; d = .75. In the 2-card condition, the mean

level of perceived responsiveness was 2.74 (SD = .80); in the 6-card condition, it was

3.27 (SD = .62).

Brief Discussion and Introduction to Study 2

Study 1 demonstrated that when same-sex strangers had a chance to become more

familiar with each other, their attraction increased. Thus, the study supports our

17

hypothesis that when familiarity is augmented in the context of social interaction, it has a

beneficial effect on attraction. Because Study 1 used an experimental design that

controlled the topics of conversation, it provides firmer evidence than earlier studies do

that familiarity may enhance attraction in social interaction. Of course, sitting in a

laboratory room discussing topics chosen by an experimenter for 30 seconds each is not

an entirely "natural" social interaction. Informational flow is restricted, participants

experience little choice about their participation or what they are saying, and the

laboratory context invites speculation about the true purposes of the experiment. Ickes

(2009) has described the often-substantial impact that such contextual factors may have

on spontaneous social interaction. Therefore, in Study 2, we sought to create a more

natural and engaging context for testing our hypothesis. We did this by randomly

assigning strangers to chat online for a varying number of instances. There were no

restrictions on the content of their chats or timing of their chats (save that we asked that

the chats last about 15 minutes each and occur within the same numbers of days in their

assigned number-of-chats condition). Additionally, participants could access the instant

messaging program from any location. Once again, we hypothesized that greater numbers

of chats would be associated with greater attraction. We also hypothesized that this

association would be mediated by perceived knowledge, responsiveness, and

comfort/satisfaction, as described earlier.

Study 2 Method

Participants

We recruited 242 participants (178 females) from the Psychology participant pool

in exchange for course extra credit. Their mean age was 19.46 (SD = 1.23), ranging from

18 to 24. Eight dyads never made contact with each other and were not considered. Three

18

additional dyads, 1 each in the 2-chat, 6-chat, and 8-chat conditions, were dropped due to

completing fewer than 50% of the required numbers of chats4. Thus study analyses were

based on 110 dyads. To encourage timely compliance with our procedures, participants

received a lottery ticket for one of six cash prizes, three for $50 and three for $25,

according to the following schedule: Having the first conversation within 24 hours = 2

lottery tickets; having no more than one conversation each day = 5 lottery tickets;

completing all conversations = 5 lottery tickets (regardless of condition); completing the

follow up surveys on time = 4 lottery tickets. Thus, participants could earn up to 16

lottery tickets, and this was unrelated to their condition or the number of chats.

Measures

Pre-Chat Measures. Students were asked to describe how much they like and feel

similar to the “average” student at the University of Rochester. These ratings were made

using 7-point Likert-type scales, anchored by 1 (Not at All) to 7 (Extremely).

Post-Chat Measures. After their final chat, participants were emailed a link to an

online survey. On average, post-chat data were obtained 1.18 days after the final chat

(SD=1.86, range, 0-12 days). Participants were asked to provide several ratings of their

partner and interactions. As in Study 1, four items contributed to a composite measure of

attraction toward the partner: perceived similarity (assessed on 7-point scales, anchored

by Not at All and Extremely), liking for the partner and desire to have the partner as a

friend (assessed on 7-point scales, anchored by Neutral (or dislike) and An exceptional

amount), and perceived closeness (the Inclusion of Other in the Self measure; Aron et al.,

1992). Values were translated to a 7-point scale, with higher numbers denoting greater

closeness. As in Study 1, these four items were combined into a composite measure of

attraction (alpha = .87).

19

Additional items assessed perceived responsiveness (12 items, based on the

measure developed by Reis, 2006, assessed with 7-point scales ranging from Not at all

true to Completely true; alpha = .95), perceived knowledge about their partner’s interests,

family, school, and social life (4 items taken from Norton et al., 2007, assessed with 7-

point scales ranging from Nothing at all to An exceptional amount; alpha = .72), their

perception of their partner's knowledge about them (identical items and scales, alpha =

.74), and satisfaction, enjoyment, and comfort with the interactions (5 items, 5-point

scales, Not at all to Extremely; alpha = .90). Participants were also asked if they wanted

to contact their partner after the study was over. We provided contact information to

partners who mutually indicated interest in contact.

Two-Week Follow-Up. Two weeks after their final conversation, participants were

sent an email asking them to indicate whether or not they had attempted to contact their

partner. Two hundred and six participants (94%) responded to this question. Two dyads

did not respond to the two-week follow-up email and 98 dyads included data from both

partners. Ten dyads included data from only one partner, and this response was used in

the analyses.

Coding of Conversations. All transcripts were coded in random order by three

independent coders who were unaware of condition, the hypotheses of the study, or the

identity of the conversants. In all five coders were used. One coder rated all transcripts,

whereas the other 4 coders each rated approximately half of the transcripts, randomly

assigned. There were a few differences between coders in their mean ratings and

variance. These differences were corrected by adjusting all scores so that each coder had

the same mean and standard deviation for each dimension coded. Coders rated each

transcript on 5 dimensions relevant to this study, using 7-point scales: knowledge and

20

familiarity about each other (“To what degree does the individual seem to know the other

person’s traits, facts about their life, etc.”), references to prior conversations (“How many

times did partners refer to things they talked about in previous conversations?”),

perceived liking (“Based on the conversation as a whole, how much do you think the

participant likes their partner?), self-disclosure of personal information (“To what extent

did the individual reveal relatively private information and emotions?”), and degree of

emotional expression (“To what extent did the participant explicitly say that they feel or

were feeling a particular emotion?”). Correlations between each pair of two coders for the

5 coded variables ranged from .45 to .64, with a mean r = .54. Scores were created by

summing the adjusted scores of all three coders for each transcript.

We then reduced these five codes to three variables by creating composite scores

for expressed knowledge (the mean of knowledge and references; r(106) = .77, p<.001)

and self-revelation (the mean of self-disclosure and emotional expression; r(106) = .38,

p<.001). The perceived liking code was used on its own.

Finally, we computed the total number of words in each chat transcript using the

Linguistic Inquiry and Word Count (LIWC) program (Pennebaker, Francis, & Booth,

2001).

Procedure

Pre-Chat Briefing. After obtaining consent, experimenters provided instructions

to participants one at a time. Participants were told that they would be paired with another

student from the university who was not affiliated with the study, whom they did not

know, and who was same sex. To minimize the chance that participants knew their

partner, we recorded their dorm, and where applicable, sports teams, Greek organization,

and other extracurricular activities that occupied at least 5 hours per week. Individuals

21

were randomly assigned to conditions, using a random numbers table, then matched with

the last unpaired person, with the constraint that individuals who matched on any of the

aforementioned criteria were not partnered. We then provided full instructions for the

study.

Participants were given an anonymous screen name and password with which to

log into the instant messaging client (e.g., urpsych1). They were asked to chat for

approximately 10 to 15 minutes with their partner for a pre-specified number of times,

depending on condition (1, 2, 4, 6, or 8 times), once a day if possible but spanning no

more than the number of days in the assigned condition (e.g., 8 days for the 8-chat

condition). We requested that participants not interact more than twice a day, but if they

did so, to space their interactions by at least two hours. We provided no guidelines on

how to structure interactions, except to request that each participant contribute to each

conversation at least 6 times (to keep it from becoming one-sided). Participants were free

to talk about whatever topics naturally arose during the course of their conversation. We

encouraged participants to maintain anonymity as follows:

"In all likelihood you won’t know this person, and in fact, our preference, for your

safety and for the integrity of the study, is that you don’t reveal your true identity

while online with your partner. Identities may come up if you choose, but our

preference is for you to remain anonymous. You will have an opportunity at the

end of the study to reveal your identity to the other person."

The chats were not pre-scheduled but were instead left to occur naturally. The

first interaction was often the most difficult to arrange, as it required both participants to

be online at the same time, using the screen names we provided rather than their own. We

used AOL-Instant Messenger (AIM), which allows individuals to post times that they

22

plan to be available to chat as an "away message." If a pair failed to connect within 24

hours, an experimenter intervened by emailing participants individually and manually

coordinating the first interaction. This occurred for 8 dyads.

After receiving instructions, participants provided the pre-chat measures, as well

as a series of self-report trait measures not relevant to the current research.

Compliance Procedure. Every day, an experimenter sent separate emails to all

active participants, reminding them to chat with their partner, and informing them about

how many conversations had been completed and how many remained. Included in daily

emails was a link for participants to follow to submit their conversations. Participants

were instructed to follow the link after every conversation and submit a transcript of their

conversation to our online survey website.

Study 2 Results

Because individual responses were nested within dyads, producing

interdependence, the primary unit of analysis was the dyad (Kenny et al., 2006), as in

Study 1. Except where otherwise noted, all variables were the average of both partners'

responses. Data were analyzed with planned contrasts, in which the primary hypothesis

specified a linear trend across the five conditions – namely, that attraction (or other

relevant variables) would increase as a function of condition. A linear trend contrast uses

1 degree of freedom (df). Rosenthal, Rosnow, and Rubin (2000) recommend that when a

study has multiple degrees of freedom but the primary hypothesis uses only 1 df, one

should test for significant effects in the remaining between-groups variance with F-

noncontrast, which in this case has 3 df (5 groups have 4 df minus 1 for the hypothesized

linear contrast). This analysis identifies whether significant nonlinear effects were

present.

23

Preliminary analyses were conducted to examine whether the linear or F-

noncontrast effects on key study variables interacted with dyad sex. No significant effects

were found. Consequently, sex was not incorporated into the analyses reported below.

We also examined pre-chat ratings of how much participants liked and felt similar to the

average University of Rochester student, to ensure that there were no pre-existing

differences across conditions. For each question, the linear contrast was non-significant,

Fs(1, 102) < 1.01, ns, as was the non-contrast effect, Fs(3, 102) < 1.30, ns.

Compliance

Compliance with our complex protocol was strong. In the 1-chat condition, 18

dyads completed 1 chat. In the 2-chat condition, 18 dyads completed 2 chats and 2 dyads

completed 1 chat. In the 4-chat condition, 21 dyads completed all 4 chats, 1 dyad

completed 5 chats, 1 dyad completed 3 chats, and 2 dyads completed 2 chats. In the 6-

chat condition, 25 dyads completed all 6 chats, 4 completed 7 chats, 1 completed 5 chats,

and 1 completed 4 chats. In the 8-chat condition, 10 dyads completed all 8 chats, 1 dyad

completed 9 chats, 3 completed 7 chats, and 2 completed 6 chats. Thus, for the analytic

sample, 92 out of 110 dyads (84%) completed the required number of chats, and 105 out

of 110 (95%) came within one of the required number.

Hypothesis Tests

Our main hypothesis was that there would be a linear increasing trend for

attraction across the 5 conditions. This hypothesis was tested by examining ratings from

the post-chat measure. Attraction produced a significant linear trend across conditions,

F(1, 101) = 5.15, p<.03. Figure 1 displays the relevant condition means. The residual

between-conditions variance was not significant, F-noncontrast (3,101) = 0.98, ns.

24

The perceived knowledge and responsiveness measures also yielded significant

linear contrasts (see Table 1 for the relevant descriptive statistics). For perceived

knowledge of the partner, the linear trend was significant, F(1, 99) = 43.85, p<.001, as

was the non-contrast test, F(3, 99) = 3.42, p<.03. Post hoc analyses indicated that this

result stemmed from the larger increase from 1-chat to 2-chat relative to the other

condition differences. Perceived knowledge of the self also produced a significant linear

trend across conditions, F(1, 99) = 36.19, p<.001, and a significant effect for the

remaining variance, F-noncontrast (3,99) = 2.76, p<.05. Post hoc tests again revealed that

this was due to a larger increase from 1-chat to 2-chat relative to the other condition.

Perceived responsiveness yielded similar results: a significant linear trend across

conditions, F(1, 101) = 11.30, p<.001, and a non-significant effect for the remaining

variance, F-noncontrast (3,101) = 1.42, ns. As Table 1 shows, all three variables showed

increases as the number of chats increased. We also analyzed how satisfied and

comfortable participants felt during their chats. As shown in Table 1, the linear trend was

significant, F(1, 101) = 7.32, p<.01, while the residual between-groups variance was not,

F-noncontrast (3,101) = 1.05, ns.

After the chats had been completed, we gave participants the opportunity to learn

each other's identity, so that they might continue their interactions. This was done in part

to provide a more externally valid measure of attraction. Two measures were obtained.

The first examined the percentage of dyads in which at least one participant expressed the

desire to learn the other's identity in order to get in contact. This was done with a one-

tailed linear χ2 test, using Fisher's exact probability, and was significant, χ2 (1) = 6.18,

p<.01. The percentages of dyads in which at least one person expressed a desire for

further contact were 17.6%, 41.2%, 52.4%, 51.6%, and 62.5%, in the 1-, 2-, 4-, 6-, and 8-

25

chat conditions, respectively. The second test, based on a question asked two weeks after

the final chat, looked at the percentage of dyads in which at least one participant reported

actually trying to contact the other. Because some participants exchanged names or other

identifying information during their conversations, we included all participants in this

analysis rather than restricting it to those for whom we had facilitated an exchange of

contact information. A similar one-tailed linear χ2 test, using Fisher's exact probability,

was significant, χ2 (1) = 2.83, p=.05. The percentages of dyads in which at least one

person actually attempted contacting the other were 11.1%, 15.0%, 25.0%, 26.7%, and

31.3%, in the 1-, 2-, 4-, 6-, and 8-chat conditions, respectively.

Thus, participants in the 8-chat condition were almost four times more likely than

participants in the single-chat condition to desire further contact after their scheduled

chats, and were almost three times more likely to actually try to contact their partner

within two weeks.

Coding of Transcripts

In order to determine whether participants' perceptions of their interactions would

be reflected in objective characteristics of those conversations, we analyzed independent

coder ratings of the last conversation that each dyad had. Here, we examined only the

final transcript for each dyad, because we wanted to determine whether the follow-up

ratings (which were obtained after the final chat had occurred) reflected the state to which

participants' relationships had evolved over their various chats. Results of these analyses

are displayed in Table 2. For each variable, the rated value generally increased as the

assigned number of chats increased. There was a strong linear trend for expressions of

knowledge about each other to increase as a function of the number of chats, F(1,104) =

25.29, p<.001. The noncontrast effect was also significant, F(3,104) = 3.56, p<.005,

26

reflecting an apparent drop-off in the 8-chat condition5. Ratings of perceived liking also

yielded significant linear trend, F(1,104) = 4.17, p<.05, and a non-significant noncontrast

effect, F(3,104) = 0.30, ns. Likewise, self-revelation yielded a significant linear trend,

F(1,104) = 5.83, p<.02, and a non-significant noncontrast effect, F(3,104) = 0.06, ns.

As shown in Table 2, there were no meaningful differences across conditions in

number of words used in these conversations, F(1,104) = 0.00, ns.

Mediation Analyses

To examine mediation of the effect of familiarity on attraction, we adopted a

bootstrapping approach (Preacher & Hayes, 2004, 2008). Bootstrapping evaluates the

magnitude of an indirect effect by comparing it to the sampling distribution of multiple

resamples of the data set. Importantly, because bootstrapping makes no assumptions

about normality in the data distribution, it corrects for possible bias in the obtained

confidence intervals, thereby improving the quality of inference in mediation models

(MacKinnon, Lockwood, & Williams, 2004). We used this method to evaluate whether

perceived knowledge, perceived responsiveness, and comfort/satisfaction mediated the

linear effect of familiarity on attraction. For simplicity, we combined the two knowledge

variables into a single composite. Contrasts representing the residual between-groups

variance were also included as control variables. A total of 5000 resamples and 95%

confidence intervals (i.e., establishing that the indirect effects were significant at two-

tailed p<.05) were used.

Because our theoretical model specified that knowledge, responsiveness, and

comfort would all play a role in mediating the effects of familiarity on attraction, we first

evaluated each of the three mediators individually. In each case, the indirect effect was

significant, and the direct effect of the linear contrast was reduced to non-significance.

27

For perceived responsiveness, the indirect effect was significant (ab =.167, SE = 0.052,

95% CI: [0.071, 0.275}; direct effect b = .006, SE = .050, t(100) = 0.13, p<.90.). For

comfort/satisfaction, the indirect effect was significant (ab =.136, SE = 0.058, 95% CI:

[0.030, 0.262}; direct effect b = .036, SE = .052, t(100) = 0.69, p<.49.). For perceived

knowledge, the indirect effect was also significant (ab =.243, SE = 0.056, 95% CI:

[0.146, 0.371}; direct effect b = -.070, SE = .070, t(98) = -1.00, p<.33.). Thus, when

considered independently, each of the three proposed mediating variables yielded a

significant indirect effect.

Because the three potential mediators were highly correlated (knowledge-

responsiveness, r[104]=.64; knowledge-comfort, r[104]=.56; responsiveness-comfort,

r[106]=.72; all ps<.001), we also evaluated our mediation hypothesis by entering all three

mediators simultaneously in a single model, following the same procedures as described

above. Figure 2 displays results from this mediation analysis. The indirect effect for

perceived responsiveness was significant (ab=.100, SE=0.038, 95% CI: {0.039, 0.189}),

as was the indirect effect for comfort/satisfaction (ab=.063, SE=0.033, 95% CI: {0.013,

0.148}). In other words, controlling for the other mediators, the linear main effect, and

the residual between-groups variance, perceived responsiveness and comfort/satisfaction

significantly mediated the effect of the contrast on attraction. For perceived knowledge,

however, the indirect effect was not significant (ab=.042, SE=0.040, 95% CI: {-0.028,

0.134}), which seems likely due to the high covariation between the three mediators. In

sum, as Figure 2 displays, when the three mediators are included in the model, the linear

familiarity contrast becomes non-significant (b=-.032, SE=0.051, t(96)=-0.63, p<.53).

When multiple mediators are included, the Preacher and Hayes bootstrapping

approach also permits comparisons of the magnitude of the indirect effects. None of these

28

contrasts was significant, indicating that in the simultaneous analyses the three indirect

effects did not differ significantly from one another in magnitude.

Brief Discussion of Study 2

As hypothesized, attraction increased between chat partners as the number of

conversations increased. Several other variables thought to be relevant to the familiarity

effect also increased correspondingly: perceived knowledge of the other, perceived

knowledge of oneself by the other, perceived responsiveness, and satisfaction/comfort

with their chats. No other trends were consistently significant across condition. Coder

ratings of perceived knowledgeability, liking, and self-revelation revealed similar trends.

Finally, two behavioroid measures of attraction, expressions of the desire to reveal one's

identity to get in contact and reports of trying to get into contact with the other, supported

findings from the self-report and coder ratings by showing increasing frequencies as a

function of the number of chats. Mediation analyses generally supported our reasoning

that responsiveness, comfort, and knowledge would account for the attraction effect,

although the indirect effect for knowledge fell short of significance when all three

mediators were included simultaneously. Thus, our hypothesis that in natural social

interaction, familiarity is associated with increased attraction was supported.

General Discussion

Across two experiments using different versions of a live interaction paradigm,

we found clear support for our hypothesis that increasing familiarity would be associated

with greater attraction. In study 1, same-sex strangers discussed for about 30 seconds

each either 2 or 6 topics chosen to facilitate the type of conversation that promotes

acquaintanceship (Aron et al., 1997; Sedikides et al., 1999). Chats in study 2 were more

spontaneous and unconstrained – conditions that allow individual differences and

29

personal preferences to emerge (Ickes, 2009). In this case, randomly paired same-sex

strangers engaged in varying numbers of online chats. Both studies found increased

attraction as a function of increasing amounts of interaction. Importantly, study 2

included measures of three variables thought to be responsible for these effects: perceived

knowledge, perceived responsiveness, and comfort/satisfaction. Linear condition effects

for these variables corresponded to the results for attraction, in both participants’ self-

reports and observer’s independent coding. Clear support was obtained for perceived

responsiveness and comfort/satisfaction as mediators of the familiarity-attraction link;

evidence also supported perceived knowledge as a mediator, although this evidence was

somewhat more tentative.

The present studies are to our knowledge the first experiments of familiarity and

attraction that directly manipulated the amount of social contact while allowing

interactions to unfold in a natural, spontaneous way. These results directly contest Norton

et al.’s (2007) conclusion that “familiarity breeds contempt” (p. 97). To be sure, we do

not dispute the validity of Norton et al.’s findings, insofar as they go. That is, if

familiarity is limited to the assessment of decontextualized positive and negative trait-

adjective descriptors, it is plausible that familiarity breeds contempt, as their results

indicate. Our position, however, is that familiarity in actual social activity is considerably

more complex and personally engaging than this. As described in the introduction, in

spontaneous social interaction, people typically share information in a more interactive,

contextualized, and individuated manner, and they process that information more

holistically (Eastwick et al., 2010). In other words, interaction partners often learn about

each other in a manner that highlights common interests and promotes enjoyable

conversation and favorable impressions. Importantly, in our study 2, both measures of

30

perceived knowledgeability – perceptions of self-knowledge about the other and of other-

knowledge about the self – were greater as a function of increased familiarity, for both

participant self-reports and observer codes. Knowledge is the exact factor that Norton et

al. (2007) theorize should decrease attraction – recall their argument that information

precludes the perception of similarity that ambiguity fosters – yet in our research both

knowledge and attraction increased as a function of familiarity, as we (and prior work on

the familiarity effect) have asserted.

Our finding is consistent with prior studies showing that more experience with

others may increase confidence and decrease uncertainty in impressions of others (both of

which seem likely to contribute to liking). For example, Swann and Gill (1997) found

that relationship length in both romantic and same-sex friendship contexts was positively

associated with confidence about impressions of others, an association that was mediated

by the richness of these impressions (a composite of the amount, integration, and

accessibility of information about the other). Similarly, Oskamp (1965) showed that

clinical psychologists' confidence about personality judgments of hypothetical cases

increased as a function of the amount of information they had available. Moreover,

Funder and his colleagues (e.g., Funder & Colvin, 1988; Letzring, Wells, & Funder,

2006) have shown that the quantity and diversity of information that friends and intimates

have about each other (in comparison to less well-acquainted others and strangers) is

associated with greater accuracy in judging personality traits, a key provision of his

Realistic Accuracy Model (Funder, 1995).

Two other mediating variables contributed to our observed effects, perceived

responsiveness and comfort/satisfaction. We think it likely that these factors are salient

when people are focused on interacting with others. Familiarity promotes responsive

31

interaction and hence attraction in several ways. First, when people become acquainted,

they are better able to offer personally relevant encouragement and support (Reis et al.,

2004). Also, familiarity fosters the provision of self-verifying feedback, because partners

are more aware of each other’s attributes and viewpoints (Swann, 1990). Finally, we

propose that there is a mutually reinforcing cycle of increasing responsiveness and

attraction, consistent with Canevello and Crocker’s (in press) findings about

responsiveness goals: The more one is attracted to the other, the more motivated one is to

be responsive to him or her. Responsive behavior by one partner, in turn, fosters the

other’s attraction. This is reminiscent of “mutual cyclical growth,” a process described by

Wieselquist, Rusbult, Foster, and Agnew (1999) with regard to the development of trust.

One partner’s pro-relationship behavior enhances the other’s satisfaction and

commitment, which in turn leads that partner to engage in pro-relationship behavior that

benefits the other. It is also consistent with Clark's account of partners' mutual concern

for each other's well-being in communal relationships (e.g., Clark & Mills, 2001).

As for the third mediator, comfort/satisfaction, a similar self-reinforcing cycle

seems like to be operative. Familiarity promotes feeling comfortable with a new

acquaintance, and as comfort increases, so does attraction. Presumably this would occur

in natural social interaction except when those interactions are aversive (Ebbesen et al.,

1976). Increased attraction may then contribute to the likelihood that subsequent

interactions would be experienced as comfortable and rewarding (Berg, 1984), adding to

the likelihood of increased attraction (e.g., Collins & Miller, 1994).

Conceptually, our results and those of Norton et al. (2007) may be reconciled by

considering the difference between evaluating information and interacting with others,

which in turn may have influenced the mindset of participants in their and our studies. As

32

described earlier, their research emphasized assessment: Participants were given bits of

information and asked how much they might like this person. Our studies, in contrast,

emphasized locomotion: Participants had to interact with peers, which would make

salient the goal of engaging in a pleasant, enjoyable interaction (Snyder & Haugen,

1995). Both orientations apply to relationship initiation and developing acquaintanceship,

albeit in different contexts. Some contexts (e.g., online dating) emphasize assessment

orientation, by providing information and asking participants to evaluate the likeability of

that person. Because assessment of static information often tends to be critical (Frost et

al., 2008; Kumashiro et al., 2007), in these instances familiarity may provide more

reasons to regard another person warily. On the other hand, spontaneous interaction

usually involves goals of one sort or another (e.g., to enjoy interacting, to accomplish

some task), which would foster a locomotive orientation. In this more natural context,

familiarity is more likely to increase attraction, for the reasons discussed earlier and as

our results demonstrate. Of course, our research did not manipulate these orientations so

we cannot say whether this proposed conceptual reconciliation of our results with Norton

et al.'s is correct. Further research is needed.

This study adds to the existing literature on familiarity in another way. Prior

studies generally used one of two formats: experimental studies in which pictures or

names were shown to participants for varying frequencies; or correlational studies in

which the frequency of social exposure or contact was measured and then linked to

attraction. (Moreland and Beach, 1992, is an exception, but note that that experiment

involved no actual interaction between participants and confederates.) As mentioned

above, the present experiments are to our knowledge the first that directly manipulated

the amount of social contact in a naturalistic context. Furthermore, note that our studies

33

controlled several other important processes that, in natural acquaintanceships, may be

confounded with increasing familiarity. For example, study 1 controlled the topic of

conversations. In study 2, participants did not meet face-to-face (and thus had no

information about physical attractiveness) or learn each other’s name until after the study

had been completed. Thus, they had no information about appearance or popularity or

what others in their social network thought about the partner. All they knew was what

they had learned and experienced through interacting. In sum, our studies add empirical

strength to the principle that familiarity fosters attraction.

It bears noting that although the three mediators we examined are conceptually

distinguishable, we cannot determine whether each one is necessary or sufficient to

produce the effects (although the analyses conducted with individual mediators suggest

that each is influential). On the other hand, in actual acquaintanceships, it seems likely

that all three processes co-occur, as they did here. We think it unlikely that

responsiveness would grow in a relationship without perceived knowledgeability or

comfort; comfort without knowledgeability and responsiveness; and knowledgeability

without responsiveness or comfort. Thus, and notwithstanding the fact that experimental

scenarios might well be able to isolate the effects of these variables relative to one

another, our analyses seem to mirror reasonably closely the emergence of attraction from

familiarity as it occurs in real-life.

Limitations

These results should be considered in the context of several limitations. First, both

studies required participants to engage in a pre-determined number of conversations to

receive course credit. This effectively ruled out the possibility of terminating the

relationship after a single, perhaps unrewarding conversation, as may often occur in real-

34

life. (Of course this limitation might also be considered a strength, in showing that

attraction can develop after artificially inducing repeated interactions with a stranger to

whom one initially feels no attraction or perhaps even repulsion; Aron, Steele, Kashdan,

& Perez, 2006). Second, because all participants were reasonably similar in certain

respects (i.e., young adults studying at the same university, taking a psychology course),

it is possible that enough pre-existing similarity existed to predispose attraction.

Similarity effects on attraction are well-established, particularly with regard to

relationship initiation (Montoya, Horton, & Kirchner, 2008). In contrast, the information

acquired in Norton et al.’s (2007) research fostered impressions of dissimilarity. Of

course, in actual social encounters, people are more likely to come into contact with

relatively similar than dissimilar others (e.g., Bandura, 1982), but it would be interesting

to see if comparable results would be obtained with a more diverse sample or when

people are explicitly paired with outgroup members (e.g., an Israeli Jew with a

Palestinian Muslim).

A third limitation is that we did not assess participants’ actual knowledge about

each other (which might have been done, for example, by comparing ratings of the other

with self-ratings). Nevertheless, existing research suggests that perceived

knowledgeability may be more important than actual knowledge for the development of

relationships (Murray, Holmes, Bellavia, Griffin, & Dolderman, 2002; Pollmann &

Finkenauer, 2009). In this regard, independent coders’ ratings confirmed participants’

self-reports of increased knowledgeability. A fourth limitation is that this research was

conducted in an individualist culture, which tends to encourage outgoing behavior and

the casual formation of social bonds. Several studies have suggested that friendships may

35

develop more cautiously in collective cultures (e.g., Triandis, Bontempo, Villareal, Asai,

& Lucca, 1988) and in such cultures, it may take longer for familiarity effects to emerge.

Conclusion

These live-interaction experiments indicate that researchers may feel confident

returning familiarity to the pantheon of social-psychological processes that promote

attraction. To be sure, there are boundary conditions to this effect and circumstances in

which it does not hold true. It would be desirable in future research to identify these

conditions, particularly contextual factors and individual differences (e.g., age,

personality) that may moderate familiarity effects. However, as a general rule, in

spontaneous, everyday social interactions among newly acquainted peers, familiarity

does indeed tend to breed liking rather than contempt. In other words, less is, after all,

less.

36

Acknowledgements

Preparation of this manuscript was partly supported by a grant from the Fetzer

Institute. We thank the following individuals for their assistance in conducting this

research: Eric Anicich, Igua Benjamin, Stacy Congdon, Avi Dressler, Malorie Elmer,

Arielle Friedlander, Beth Kacel, Molly Korotkin, Jacquie Kramer, Sarah Mancuso,

Aquella Robinson, Ali Rosenbluth, Samantha Siskind, Penelope Spicer, Max Sutton-

Smolin, and Jo Ellyn Walker.

37

Footnotes

1 One important methodological limitation of Norton et al.'s study 5 is that all

participants were recruited from an online dating website, so everyone who provided

post-date ratings had gone on a date, but then returned to the dating website without an

explicit invitation by researchers to do so. Presumably, this means that the date had not

gone well, since if the date had gone well, there would be lesser motivation to return to

the dating website to look for a new date.

2 Two of Norton et al.'s studies were run in a dating context and presumably involved

opposite-sex pairs. The other two studies were not conducted in a dating context, and the

sex-pairing is not mentioned in their article.

3 If data from both of the omitted participants are used, the pattern of results remains the

same.

4 Findings remain unchanged if these dyads are included.

5 Consistent with this observation, the quadratic trend was significant, F(1,104) = 9.76,

p<.005.

38

References

Amir, Y. (1976). The role of inter-group contact in change of prejudice and ethic

relations. In P.A. Katz (Ed.), Towards the elimination of racism (pp. 245-308).

New York: Pergamon.

Aron, A., Aron, E. N., & Smollan, D. (1992). Inclusion of other in the self scale and the

structure of interpersonal closeness. Journal of Personality and Social

Psychology, 63, 596-612.

Aron, A., Melinat, E., Aron, E. N., Vallone, R. D., & Bator, R. J. (1997). The

experimental generation of interpersonal closeness: A procedure and some

preliminary findings. Personality and Social Psychology Bulletin, 23, 363-377.

Aron, A., Steele, J. L., Kashdan, T. B., & Perez, M. (2006). When similars do not attract:

Tests of a prediction from the self-expansion model. Personal Relationships, 13,

387–396.

Asch, S. E. (1946). Forming impressions of personality. Journal of Abnormal and Social


Back, M. D., Schmukle, S. C., & Egloff, B. (2008). Becoming friends by chance.

Psychological Science, 19, 439-440.

Bandura, A. (1982). Psychology of chance encounters and life paths. American

Psychologist, 37, 747-755.

Bargh, J. A., Gollwitzer, P. M., & Oettingen, G. (2010). Motivation. In S. Fiske, D. Gilbert,

& G. Lindzey (Eds.), Handbook of Social Psychology (5th ed., pp. 268-316). New

York: Wiley.

39

Baumeister, R. F., & Leary, M. R. (1995). The need to belong: Desire for interpersonal

attachment as a fundamental human motivation. Psychological Bulletin, 117, 497-

529.

Berg, J. H. (1984). Development of friendship between roommates. Journal of

Personality and Social Psychology, 46, 346-356.

Berscheid, E., & Ammazzalorso, H. (2001). Emotional experience in close relationships. In

G. J. O. Fletcher & M. S. Clark (Eds.), Blackwell handbook of social psychology

(Vol. 2, pp. 308-330). Oxford, UK: Blackwell.

Berscheid, E., & Regan, P. (2005). The psychology of interpersonal relationships. Upper

Saddle River, NJ: Pearson Education.

Berscheid, E., & Walster, E. H. (1969). Interpersonal attraction. Reading, MA: Addison

Wesley.

Bornstein, R. F. (1989). Exposure and affect: Overview and meta-analysis of research, 1968-

1987. Psychological Bulletin, 106, 265-289.

Bornstein, R. F., & D'Agostino, P. R. (1994). The attribution and discounting of perceptual

fluency: Preliminary tests of a perceptual fluency/attributional model of the mere

exposure effect. Social Cognition, 12, 103-128.

Brickman, P., Meyer, P., & Fredd, S. (1975). Effects of varying exposure to another person

with familiar or unfamiliar thought processes. Journal of Experimental Social


Brockner, J., & Swap, W. C. (1976). Effects of repeated exposure and attitudinal similarity

on self disclosure and interpersonal attraction. Journal of Personality and Social


40

Canevello, A., & Crocker, J. (in press). Creating good relationships: Responsiveness,

relationship quality, and interpersonal goals. Journal of Personality and Social

Psychology.

Clark, H. H. (1996). Using language. Cambridge: Cambridge University Press.

Clark, H. H., & Brennan, S. A. (1991). Grounding in communication. In L.B. Resnick,

J.M. Levine, & S.D. Teasley (Eds.). Perspectives on socially shared cognition

(pp. 127-149). Washington, DC: APA Books.

Clark, L. A., & Watson, D. (1988). Mood and the mundane: Relations between daily life

events and self-reported mood. Journal of Personality and Social Psychology, 54,

294-306.

Clark, M. S., & Mills, J. (2001). Behaving in such a way as to maintain and enhance

relationship satisfaction. In J. H. Harvey & A. E. Wenzel (Eds.), Relationship

maintenance and enhancement (pp. 13-26). Mahwah, NJ: Erlbaum.

Collins, N. L., & Miller, L. C. (1994). Self-disclosure and liking: A meta-analytic review.

Psychological Bulletin, 116, 457-475.

Curtis, R. C., & Miller, K. (1986). Believing another likes or dislikes you: Behaviors making

the beliefs come true. Journal of Personality and Social Psychology, 51, 284-290.

Denrell, J. (2005). Why most people disapprove of me: Experience sampling in

impression formation. Psychological Review, 112, 951-978.

Druckman, D., & Broome, B. J. (1991). Value differences and conflict resolution:

Familiarity or liking? Journal of Conflict Resolution, 35, 571-593.

Eastwick, P. W., & Finkel, E. J. (2008). Sex differences in mate preferences revisited: Do

people know what they initially desire in a romantic partner? Journal of


41

Eastwick, P. W., Finkel, E. J., & Eagly, A. H. (2010). When and why do people fail to

use their ideal partner preferences when selecting and retaining a romantic

partner? Manuscript under editorial review.

Ebbsen, E. B., Kjos, G. L., & Konecni, V. J. (1976). Spatial ecology: Its effects on the

choice of friends and enemies. Journal of Experimental Social Psychology, 12,

505-518.

Festinger, L., Schachter, S., & Back, K. (1950). Social pressures in informal groups: A study

of human factors in housing. Stanford, CA: Stanford University Press.

Floyd, K. (2006). Communicating affection. New York: Cambridge University Press.

Frost, J. H., Chance, Z., Norton, M. I., & Ariely, D. (2008). People are experience goods:

Improving online dating with virtual dates. Journal of Interactive Marketing, 22,

51–62.

Funder, D. C. (1995). On the accuracy of personality judgment: A realistic approach.

Psychological Review, 102, 652-670.

Funder, D. C., & Colvin, C. R. (1988). Friends and strangers: Acquaintanceship, agreement,

and the accuracy of personality judgment. Journal of Personality and Social


Gollwitzer, P. M., & Bayer, U. (1999). Deliberative versus implemental mindsets in the

control of action. In S. Chaiken & Y. Trope (Eds.), Dual-process theories in social

psychology (pp. 403-422). New York: Guilford.

Hardin, C. D., & Higgins, E. T. (1996). Shared reality: How social verification makes the

subjective objective. In R. M. Sorrentino & E. T. Higgins (Eds.), Handbook of

motivation and cognition (pp. 28-84). New York: Guilford Press.

Harrison, A. A. (1969). Exposure and popularity. Journal of Personality, 37, 359-377.

42

Heine, S. J., & Renshaw, K. (2002). Interjudge agreement, self-enhancement, and liking:

Cross-cultural divergences. Personality and Social Psychology Bulletin, 28, 578-

587.

Ickes, W. (2009). Strangers in a strange lab: How personality shapes our initial

encounters with others. New York: Oxford University Press.

Kelley, H. H., Berscheid, E., Christensen, A., Harvey, J. H., Huston, T. L., Levinger, G.

McClintock, E., Peplau, L. A., & Peterson, D. R. (1983). Close relationships. New

York: Freeman.

Kenny, D. A. (1994). Interpersonal perception: A social relations analysis. New York:

Guilford Press.

Kenny, D. A., Kashy, D. A., & Cook, W. L. (2006). Dyadic data analysis. New York:

Guilford Press.

Kruglanski, A. W., Thompson, E. P., Higgins, E. T., Atash, M. N., Pierro, A., Shah, J. Y.,

& Spiegel, S. (2000). To ‘do the right thing’ or to ‘just do it’: Locomotion and

assessment as distinct self-regulatory imperatives. Journal of Personality and

Social Psychology, 79, 793–815.

Kumashiro, M., Rusbult, C. E., Finkenauer, C., & Stocker, S. (2007). To think or to do:

Assessment orientation, locomotion orientation, and the Michelangelo

Phenomenon. Journal of Social and Personal Relationships, 24, 591-611.

Lee, A. Y. (2001). The mere exposure effect: An uncertainty reduction explanation

revisited. Personality and Social Psychology Bulletin, 27, 1255-1266.

Letzring, T. D., Wells, S. M., & Funder, D. C. (2006). Quantity and quality of available

information affect the realistic accuracy of personality judgment. Journal of


43

Mikulincer, M., & Shaver, P. (2007). Attachment patterns in adulthood: Structure,

dynamics, and change. New York: Guilford Press.

MacKinnon, D. P., Lockwood, C. M., & Williams, J. (2004). Confidence limits for the

indirect effect: Distribution of the product and resampling methods. Multivariate

Behavioral Research, 39, 99-128.

Miller, L. C., & Read, S. J. (1987). Why am I telling you this? Self-disclosure in a goal-based

model of personality. In V. J. Derlega & J. H. Berg (Eds.), Self-disclosure: Theory,

research, and therapy (pp. 35-58). New York: Plenum.

Mita, T. H., Dermer, M., & Knight, J. (1977). Reversed facial images and the mere-

exposure hypothesis. Journal of Personality and Social Psychology, 35, 597-601.

Montoya, R. M., Horton, R. S., & Kirchner, J. (2008). Is actual similarity necessary for

attraction? A meta-analysis of actual and perceived similarity. Journal of Social

and Personal Relationships, 25, 889-922.

Moreland, R. L., & Beach, S. R. (1992). Exposure effects in the classroom: The

development of affinity among students. Journal of Experimental Social


Moreland, R. L., & Zajonc, R. B. (1982). Exposure effects in person perception:

Familiarity, similarity, and attraction. Journal of Experimental Social Psychology,

18, 395-415.

Murray, S. L., Holmes, J. G., Bellavia, G., Griffin, D. W., & Dolderman, D. (2002).

Kindred spirits? The benefits of egocentrism in close relationships. Journal of


Nahemow, L., & Lawton, M.P. (1975). Similarity and propinquity in friendship

formation. Journal of Personality and Social Psychology, 32, 205–213.

44

Newcomb, T. M. (1961). The acquaintance process. New York: Holt, Rinehart

&Winston.

Norton, M. I., Frost, J. H., & Ariely, D. (2007). Less is more: The lure of ambiguity, or

why familiarity breeds contempt. Journal of Personality and Social Psychology,

92, 97–105.

Orive, R., & Gerard, H. B. (1987). The familiar stimulus as a reducer of anxiety: An

experimental study. Journal of Social and Clinical Psychology, 5, 330-338.

Oskamp, Stuart (1965). Overconfidence in case-study judgements. Journal of Consulting

Psychology, 2, 261–265.

Parker-Rees, R. (2007). Liking to be liked: Imitation, familiarity and pedagogy in the first

years of life. Early Years: An International Journal of Research and

Development, 27, 3-17.

Pennebaker, J. W., Francis, R. J., & Booth, M. E. (2001). Linguistic inquiry and word

count (LIWC). Mahwah, N.J.: Erlbaum.

Perlman, D., & Oskamp, S. (1971). The effects of picture content and exposure frequency

on evaluations of Negroes and whites. Journal of Experimental Social


Pettigrew, T. F. (1998). Intergroup contact theory. Annual Review of Psychology, 49, 65-

85.

Pollmann, M. M. H., & Finkenauer, C. (2009). Investigating the role of two types of

understanding in relationship well-being: Understanding is more important than

knowledge. Personality and Social Psychology Bulletin, 35, 1512-1527.

45

Preacher, K. J., & Hayes, A. F. (2004). SPSS and SAS procedures for estimating indirect

effects in simple mediation models. Behavior Research Methods, Instruments, &

Computers, 36, 717-731.

Preacher, K. J., & Hayes, A. F. (2008). Asymptotic and resampling strategies for

assessing and comparing indirect effects in multiple mediator models. Behavior

Research Methods, 40, 879-891.

Rajecki, D. W. (1985). Predictability and control in relationships: A perspective from

animal behavior. In W. Ickes (Ed.), Compatible and incompatible relationships (pp.

11-31). New York: Springer-Verlag.

Reis, H. T. (2006). The perceived responsiveness scale. Unpublished manuscript,

University of Rochester.

Reis, H.T., & Aron, A. (2008). Love: What is it, why does it matter, and how does it

operate? Perspectives on Psychological Science, 3, 80-86.

Reis, H. T., Clark, M. S., & Holmes, J. G. (2004). Perceived partner responsiveness as an

organizing construct in the study of intimacy and closeness. In D. J. Mashek and

A. Aron (Ed.), Handbook of closeness and intimacy (pp. 201-225). Mahwah, N.J.:

Lawrence Erlbaum.

Reis, H. T., & Gable, S. L. (2003). Toward a positive psychology of relationships. In C. L.

Keyes & J. Haidt (Eds.), Flourishing: The positive person and the good life.

Washington, D.C.: APA Press.

Reis, H. T., & Patrick, B. C. (1996). Attachment and intimacy: Component processes. In A.

Kruglanski & E. T. Higgins (Eds.), Social psychology: Handbook of basic principles

(pp. 523-563). New York: Guilford Press.

46

Reis, H. T., & Shaver, P. (1988). Intimacy as an interpersonal process. In S. W. Duck (Ed.),

Handbook of personal relationships, (pp. 367-389). Chichester, UK: John Wiley and

Sons, Ltd.

Rosenthal, R., Rosnow, R. L., & Rubin, D. B. (2000). Contrasts and effect sizes in

behavioral research: A correlational approach. New York: Cambridge

University Press.

Sedikides, C., Campbell, W. K., Reeder, G., & Elliot, A. J. (1999). The Relationship

Closeness Induction Test. Representative Research in Social Psychology, 23, 1-4.

Segal, M. W. (1974). Alphabet and attraction: An unobtrusive measure of the effect of

propinquity in a field setting. Journal of Personality and Social Psychology, 30, 654-

657.

Shook, N. J., & Fazio, R. H. (2008). Interracial roommate relationships: An experimental

field test of the contact hypothesis. Psychological Science,19, 717-723.

Smith, E. R., Miller, D. A., Maitner, A. T., Crump, S. A., Garcia-Marques, T., & Mackie,

D. M. (2006). Familiarity can increase stereotyping. Journal of Experimental

Social Psychology, 42, 471-478.

Snyder, M., & Haugen, J. (1995). Why does behavioral confirmation occur? A functional

perspective on the role of the target. Personality and Social Psychology Bulletin,

21, 963-974.

Sprecher, S., Schwartz, P., Harvey, J., & Hatfield, E. (2008). The BusinessOfLove.com:

Relationship initiation at Internet matchmaking services. In S. Sprecher, A,

Wenzel, & J. Harvey (Eds.), Handbook of relationship initiation (pp. 249-265).

New York: Psychology Press.

47

Stasser, G., Taylor, L. A., & Hanna, C. (1989). Information sampling in structured and

unstructured discussions of three and six-person groups. Journal of Personality and

Social Psychology, 57, 67-78.

Swann, W. B., Jr. (1990). To be adored or to be known: The interplay of self-enhancement

and self-verification. In R. Sorrentino & E. T. Higgins (Eds.), Handbook of

motivation and cognition (Vol. 2, pp.408-448). New York: Guilford Press.

Swann, W. B., Jr. & Gill, M.J. (1997). Confidence and accuracy in person perception: Do we

know what we think we know about our relationship partners? Journal of Personality

and Social Psychology, 73, 747-757.

Triandis, H. C., Bontempo, R., Villareal, M. J., Asai, M., & Lucca, N. (1988). Individualism

and collectivism: Cross-cultural perspectives on self-ingroup relationships. Journal of


Walster, E., Aronson, V., Abrahams, D., & Rottman, L. (1966). Importance of physical

attractiveness in dating behavior. Journal of Personality and Social Psychology, 4,

508-516.

Winkielman, P., Schwarz, N., Fazendiero, T., & Reber, R. (2003). The hedonic marking

of processing fluency: Implications for evaluative judgment. In J. Musch & K. C.

Klauer (Eds.), The psychology of evaluation: Affective processes in cognition and

emotion (pp. 189-217). Mahwah, N.J.: Erlbaum.

Wolf, A. P. (1995). Sexual attraction and childhood association: A Chinese brief for

Edward Westermarck. Stanford, CA: Stanford University Press.

Zajonc, R. B. (1965). Social facilitation. Science, 149, 269-274.

Zajonc, R. B. (1968). Attitudinal effects of mere exposure. Journal of Personality and Social

Psychology Monograph Supplement, 9, 1-27.

48

Zajonc, R. B. (2001). Mere exposure: A gateway to the subliminal. Current Directions in

Psychological Science, 10, 224-228.

49

Table 1

Levels of Perceived Knowledge and Responsiveness as a

Function of the Number of Chats

Condition (Number of Chats)

1 2 4 6 8 Linear F F-noncontrast

Perceived knowledge of other

1.99 (.56)

2.92 (.56)

3.11 (.64)

3.16 (.62)

3.52 (.86) 43.85**** 3.42**

Perceived knowledge by other

1.92 (.54)

2.79 (.51)

2.97 (.78)

3.08 (.69)

3.30 (.75) 36.19**** 2.76*

Perceived responsiveness

2.40 (.72)

3.09 (.76)

3.12 (.80)

3.44 (1.08)

3.29 (.91) 11.30**** 1.42

Satisfaction/Comfort 3.11 (.78)

3.62 (.40)

3.58 (.63)

3.66 (.70)

3.77 (.71) 7.32*** 1.05

Note: Mean levels are presented; standard deviations are reported in parentheses. **** p<.001 *** p<.01 ** p<.02 * p<.05

50

Table 2

Levels of Coded Knowledge, Liking, and Self-Revelation in Dyads' Final Chat

as a Function of the Number of Chats

Condition (Number of Chats)

1 2 4 6 8 Linear F F-noncontrast

Knowledge about each other

-0.61 (.11)

-0.09 (.59)

0.42 (.69)

0.50 (.82)

0.31 (.83) 25.29*** 3.56**

Rated liking -0.14 (.51)

0.09 (.65)

0.23 (.81)

0.40 (.98)

0.31 (1.06) 4.17* 0.30

Self-revelation -0.10 (.34)

-0.04 (.39)

0.12 (.71)

0.18 (.55)

0.33 (.88) 5.83** 0.06

Word count 530.7 (159.1)

527.2 (246.6)

490.5 (142.5)

528.9 (233.0)

523.8 (239.2) 0.00 0.21

Note: Mean levels are listed; standard deviations are reported in parentheses. *** p<.001 ** p<.02 * p<.05

51

Figure Captions

Figure 1. Attraction as a function of the number of chats (condition)

Figure 2. Proposed mediational model. All coefficients are significant at p<.05 or less

except for the direct effect of the linear condition contrast when the three mediators are

included in the model.

52

(Figure 1)

2.8

3

3.2

3.4

3.6

3.8

4

1 2 4 6 8Number of Chats (Condition)

53

(Figure 2)

.21

Linear Chats

Contrast

Perceived Knowledge

Perceived Responsiveness

Comfort & Satisfaction

Attraction

.32

.14

.1

.47

.46

-.03 (.17 )

54

Appendix A

Twelve-Item Perceived Responsiveness Scale

Compared to most experiences I’ve had meeting somebody new, I get the feeling that this person: 1 2 3 4 5 6 7 not at all somewhat very completely true true true true _____ 1. ... sees the “real” me. _____ 2. ... “gets the facts right” about me. _____ 3. ... esteems me, shortcomings and all. _____ 4. ... knows me well. _____ 5. ... values and respects the whole package that is the “real” me. _____ 6. ... understands me. _____ 7. ... really listens to me. _____ 8. ... expresses liking and encouragement for me. _____ 9. ... seems interested in what I am thinking and feeling. _____ 10. ... values my abilities and opinions. _____ 11. ... is on “the same wavelength” with me. _____ 12. ... is responsive to my needs.

Date post:	14-Oct-2020
Category:	Documents
Upload:	others
View:	0 times
Download:	0 times

Familiarity Does Indeed Promote Attraction in Live Interaction … · 2011. 4. 3. · Familiarity...

Documents